summaryrefslogtreecommitdiff
path: root/text_recognizer/data/iam_paragraphs.py
diff options
context:
space:
mode:
Diffstat (limited to 'text_recognizer/data/iam_paragraphs.py')
-rw-r--r--text_recognizer/data/iam_paragraphs.py25
1 files changed, 22 insertions, 3 deletions
diff --git a/text_recognizer/data/iam_paragraphs.py b/text_recognizer/data/iam_paragraphs.py
index 9c75129..3bf28ff 100644
--- a/text_recognizer/data/iam_paragraphs.py
+++ b/text_recognizer/data/iam_paragraphs.py
@@ -1,7 +1,7 @@
"""IAM Paragraphs Dataset class."""
import json
from pathlib import Path
-from typing import Dict, List, Optional, Sequence, Tuple
+from typing import Callable, Dict, List, Optional, Sequence, Tuple
from loguru import logger as log
import numpy as np
@@ -35,8 +35,27 @@ MAX_WORD_PIECE_LENGTH = 451
class IAMParagraphs(BaseDataModule):
"""IAM handwriting database paragraphs."""
- def __init__(self) -> None:
- super().__init__()
+ def __init__(
+ self,
+ mapping: EmnistMapping,
+ transform: Optional[Callable] = None,
+ test_transform: Optional[Callable] = None,
+ target_transform: Optional[Callable] = None,
+ train_fraction: float = 0.8,
+ batch_size: int = 16,
+ num_workers: int = 0,
+ pin_memory: bool = True,
+ ) -> None:
+ super().__init__(
+ mapping,
+ transform,
+ test_transform,
+ target_transform,
+ train_fraction,
+ batch_size,
+ num_workers,
+ pin_memory,
+ )
self.dims = (1, IMAGE_HEIGHT, IMAGE_WIDTH)
self.output_dims = (MAX_LABEL_LENGTH, 1)